Systematic reviews: New challenges relating to non-randomised studies and diagnostic test accuracy

Systematic reviews of the effects of healthcare interventions are now quite common. There are currently more than 2 600 full Cochrane reviews in The Cochrane Library, with protocols published for 1 600 more. There are also thousands of systematic reviews published in other journals. However, the science of systematic reviewing is still relatively young. Most of the reviews available today rely on randomised trials, but there are also some reviews of non-randomised trials and of diagnostic test accuracy and these may become more common in the next few years. In this essay, I discuss some of the challenges of doing these newer types of systematic review, and show how experience gained in the last few decades of systematic reviews of randomised trials might help to meet these challenges